This invention relates, in general, to electronics, and more particularly, to methods of manufacturing sensors.
Many absolute pressure sensors use two parallel capacitive plates to measure a pressure where a higher pressure deflects a movable capacitive plate closer to a stationary capacitive plate. However, these sensors suffer from a non-linearity error where the movable capacitive plate is deformed in a non-l near manner by the measured pressure such that the movable capacitive plate is no longer substantially parallel to the stationary capacitive plate. The resulting non-linearity error produces an inaccurate measurement of the pressure.
xe2x80x9cOne technique to reduce the non-linearity error is to increase the thickness of the movable capacitive plate or to place additional support layers over or under the movable capacitive plate. However, the increased thickness of the movable plate produces problems when integrating the sensor onto a complimentary metal-oxide-semiconductor (CMOS) chip.xe2x80x9d
Another technique to reduce the non-linearity error is to reduce the size of the movable capacitive plate. However, the reduced size decreases the capacitance, which reduces the sensitivity of the sensor. To compensate for the reduced capacitance, an array of small, independently movable capacitive plates can be used where each of the movable capacitive plates is anchored to a support substrate that also supports the stationary capacitive plate. However, the parasitic capacitance of the sensor is significantly magnified by the increased number of anchors in the array of sensors, and this increased parasitic capacitance produces other problems in measuring the pressure.
Still another technique to reduce the non-linearity error is to use additional circuitry to compensate for the error. However, the additional circuitry increases the size and cost of the completed sensor.
Yet another technique to reduce the non-linearity error allows the movable capacitive plate to xe2x80x9ctouch downxe2x80x9d or contact the stationary capacitive plate. While this technique has been reported to produce more linear electrical characteristics, this technique suffers from a hysteresis effect. Moreover, the movable capacitive plate in this xe2x80x9ctouch downxe2x80x9d technique suffers from material fatigue problems.
Accordingly, a need exists for a method of manufacturing a sensor that can eliminate, or at least reduce, the magnitude of the non-linearity error while the sensor remains integratable on-chip with control circuitry, while maintaining the sensitivity of the sensor, while minimizing the parasitic capacitance, while maintaining reliability, without introducing hysteresis effects or material fatigue problems, and without increasing the cost of the sensor component by requiring additional compensation circuitry.